Driving is critical to independence and quality of life; ensuring safety of drivers and other road users is a major public health issue. In 2007, at least 22% of fatal accidents occurred at intersections with older drivers over represented. This is a growing concern as older drivers are the most rapidly increasing segment of the driving population. An important question in understanding the cause of a crash is whether or not the driver looked (scanned) before entering the intersection and if he did look, whether he detected the collision object. Crossroad and T-intersections are particularly challenging as a wide field (up to 180) has to be checked for potential hazards, requiring head as well as eye movements. After a crash, drivers frequently report that they looked but failed to see (LBFTS) the other vehicle/pedestrian; however, there is usually no objective measure (such as head/eye tracking) of whether the driver did indeed look. Both older age and visual field loss could increase the likelihood of either not scanning in the direction of a hazard or not scanning far enough. Efficient scanning is critical for drivers with visual field loss; if they do not scan, then a hazad that appears within a field loss area might never be seen. The overall goals of this research are to (1) address the lack of information about the extent to which inadequate scanning by older drivers and drivers with visual field loss may contribute to intersection collisions and (2) develo and evaluate a novel head-borne scanning device to address intersection scanning deficits. The device will have two applications: as a driver assistance system and as a training tool. It will monitor head scanning in the critical period before entering an intersection and trigger alerting auditory cues, when necessary, to remind the driver to scan in a specific direction. To address the first goal, intersection scenarios with realistic hazards will be designed and a systematic method of identifying LBFTS events based on a novel test will be developed. The effects of age and hemianopia (the loss of half the field of vision in both eyes) on scanning and detection will then be evaluated within the safe, controlled environment of a wide-field driving simulator. Eye and head tracking will be used to establish the relationship between scanning behaviors and detection failures and to identify LBFTS events. The hypothesis that detection failures are more frequently a result of scanning deficits than LBFTS incidents will be tested. To address the second goal, two studies will be conducted in the driving simulator to evaluate the safety and efficacy of the scanning device both as an assistance system and as a training tool for drivers with hemianopia and older drivers with normal sight. The effects of the device on intersection scanning and detection of hazards will be evaluated. The study of drivers with normal sight will also include recordings of scanning behaviors during an extended period of habitual driving before and after using the device in the driving simulator, which will provide a real-world assessment of the potential of the device as a training tool.